microsoft/DeepSpeed

Pull Requests Commits

ZeRO3: Improve mismatch detection (#7525)

sfc-gh-truwase committed 253 days ago

Verified eabb687a

Add riscv64 cpu support in deepspeed_shm_comm op (#7519)

heyujiao99 committed 255 days ago

Verified 9bf215d2

Update README with ZenFlow release blog featured by PyTorch. (#7520)

Antlera committed 256 days ago

Verified e04fa3e6

Enable non-ZeRO mode (#7515)

sfc-gh-truwase committed 257 days ago

Verified 889f0ead

Enabling Muon Optimizer in DeepSpeed (#7509)

PKUWZP committed 258 days ago

Verified 66ad2780

Update TSC Committers (#7517)

PKUWZP committed 258 days ago

Verified e4662faf

Fix assert when 'pp_int' object has no attribute 'custom_print_str' (#7507)

aeeeeeep committed 259 days ago

Verified 38d1a9eb

CI funding shout out to modal.com (#7503)

stas00 committed 263 days ago

Verified d9cb7868

Support DeepSpeed offload and reload states with ZeRO1 and ZeRO2 (#7421)

LYMDLUT committed 264 days ago

Verified bc8c0db3

Update version.txt after 0.17.5 release (#7502)

loadams committed 264 days ago

Verified f45159e4

Add index to HPU devices (#7497)

deepcharm committed 266 days ago

Verified 047a7599

Reduce performance impact of compiler.enable decorator (#7498)

deepcharm committed 266 days ago

Verified 8cf5fc57

Fix DeepCompile for PyTorch v2.8 (#7496)

tohtana committed 266 days ago

Verified 12b4dc19

Fix invalid f-strings (#7457)

cyyever committed 268 days ago

Verified 1c03d1b1

Add Zenflow code for Stage 1 & 2 (#7391)

Antlera committed 269 days ago

Verified 1d7b90ad

fix xpu device_id AttributeError issue (#7488)

yao-matrix committed 269 days ago

Verified 33cd9450

Enable forked PRs (#7486)

sfc-gh-truwase committed 270 days ago

Verified 64ac13f7

Fix pre-compile on cpu-only machines (#7168)

AlongWY committed 272 days ago

Verified 8aadf6cb

[TiledFusedLogitsLoss] support inference (#7477)

stas00 committed 273 days ago

Verified a54c3943

[UlyssesSPDataLoaderAdapter] fix iterator reset (#7472)

stas00 committed 273 days ago

Verified d75196a0

Modal CI (#7289)

sfc-gh-truwase committed 273 days ago

Verified a12de38d

fix `deepspeed --venv_script` (#7469)

stas00 committed 273 days ago

Verified 8e029923

Fix cpu CI (#7481)

sfc-gh-truwase committed 273 days ago

Verified 8c83e42b

Add blog for ZenFlow (#7463)

Antlera committed 274 days ago

Verified cda3f962

add --bind_cores_to_rank to zero offload tutorial (#7474)

delock committed 276 days ago

Verified f03d416e

fix #7188 (#7371)

lpnpcs committed 280 days ago

Verified f897b673

Fix all-gather duplicate params and wrong dtype (#7462)

eternalNight committed 282 days ago

Verified 0aff6b2c

fix issues raised by Coverity scans (#7431)

NirSonnenschein committed 282 days ago

Verified 1a8ad24f

Add getter APIs for TP/PP/DP ranks in DeepSpeedEngine (#7427)

WoosungMyung committed 283 days ago

Verified 0e51e093

Update README.md (#7465)

PKUWZP committed 283 days ago

Verified e1560d84

Newer Older